Sea Level Estimation Based on GNSS Dual-Frequency Carrier Phase Linear Combinations and SNR

• Published in: Remote sensing (Basel, Switzerland) Vol. 10; no. 3; p. 470
• Main Authors: Wang, Nazi, Xu, Tianhe, Gao, Fan, Xu, Guochang
• Format: Journal Article
• Language: English
• Published: MDPI AG 01.03.2018
• Subjects: carrier phase geometry-free linear combinations; GNSS reflectometry; sea level; SNR; tide gauge
• ISSN: 2072-4292; 2072-4292
• DOI: 10.3390/rs10030470

Ground-based GNSS-R (global navigation satellite system reflectometry) can provide the absolute vertical distance from a GNSS antenna to the reflective surface of the ocean in a common height reference frame, given that vertical crustal motion at a GNSS station can be determined using direct GNSS signals. This technique offers the advantage of enabling ground-based sea level measurements to be more accurately determined compared with traditional tide gauges. Sea level changes can be retrieved from multipath effects on GNSS, which is caused by interference of the GNSS L-band microwave signals (directly from satellites) with reflections from the environment that occur before reaching the antenna. Most of the GNSS observation types, such as pseudo-range, carrier-phase and signal-to-noise ratio (SNR), suffer from this multipath effect. In this paper, sea level altimetry determinations are presented for the first time based on geometry-free linear combinations of the carrier phase at low elevation angles from a fixed global positioning system (GPS) station. The precision of the altimetry solutions are similar to those derived from GNSS SNR data. There are different types of observation and reflector height retrieval methods used in the data processing, and to analyze the performance of the different methods, five sea level determination strategies are adopted. The solutions from the five strategies are compared with tide gauge measurements near the GPS station, and the results show that sea level changes determined from GPS SNR and carrier phase combinations for the five strategies show good agreement (correlation coefficient of 0.97–0.98 and root-mean-square error values of <0.2 m).